Method for Detecting of Clone Areas in a Digital Image under Conditions of Additional Attacks

摘要

The presented new method for detecting of clone areas in a digital image is effective in conditions of disturbing effects on the image that are additional to the cloning, including cases of a small area of clone area (less than 0.4% of the image). The method is based on the detection of geometric comparability of surface parts, which is compared with digital image, corresponding to areas of the clone and its prototype. The quantitative index was found. This index is characterizing the individual areas of the image and coinciding for areas of clone and its prototype, including under attack conditions. This index is the value of the local (global) minimum of function values that interpolate the elements of the matrix G, which is put into correspondence to the analyzed digital image matrix. The elements G represent the smallest difference between the corresponding block of the digital images matrix q x q from any other of its block. The results of the computational experiment are presented. They are confirmed high efficiency of this method in compare to the existing analogues, regardless of the specifics of attack conducted on cloned image, the relative sizes of the cloned area, block sizes used in the image analysis.